def create_inputs(data):
"""Index input reads and prepare groups of reads to process concurrently.
Allows parallelization of alignment beyond processors available on a single
machine. Prepares a bgzip and grabix indexed file for retrieving sections
of files.
"""
from bcbio.pipeline import sample
data = cwlutils.normalize_missing(data)
aligner = tz.get_in(("config", "algorithm", "aligner"), data)
# CRAM files must be converted to bgzipped fastq, unless not aligning.
# Also need to prep and download remote files.
if not ("files" in data and data["files"] and aligner and (_is_cram_input(data["files"]) or
objectstore.is_remote(data["files"][0]))):
# skip indexing on samples without input files or not doing alignment
if ("files" not in data or not data["files"] or data["files"][0] is None or not aligner):
return [[data]]
data["files_orig"] = data["files"]
data["files"] = _prep_fastq_inputs(data["files"], data)
# preparation converts illumina into sanger format
data["config"]["algorithm"]["quality_format"] = "standard"
# Handle any necessary trimming
data = utils.to_single_data(sample.trim_sample(data)[0])
_prep_grabix_indexes(data["files"], data)
data = _set_align_split_size(data)
out = []
if tz.get_in(["config", "algorithm", "align_split_size"], data):
splits = _find_read_splits(data["files"][0], data["config"]["algorithm"]["align_split_size"])
for split in splits:
cur_data = copy.deepcopy(data)
cur_data["align_split"] = split
out.append([cur_data])
else:
out.append([data])
if "output_cwl_keys" in data:
out = cwlutils.samples_to_records([utils.to_single_data(x) for x in out],
["files", "align_split", "config__algorithm__quality_format"])
return out
def trim_sample(*args):
return sample.trim_sample(*args)
def trim_sample(*args):
return sample.trim_sample(*args)
